Vehicle Fuzzy Pattern Recognition In Intelligent Transportation Systems

With the development of economy, the demand of traffic enhanced continuously, while the traffic management is relatively lag, which makes the traffic jam increasingly serious. One of way to improve it is the introduction of intelligent transportation system. In this paper, we did some research on pattern recognition of vehicle automatic in the intelligent transportation system.We screen the vehicle images from just over and just side, getting the characteristics of vehicle after pretreatment, partition, profile tracking. We introduce fuzzy degree for character choice and using fuzzy degree as index of character for contribution of classification. Because there are many kinds of vehicle in our country with no distinct difference among them, it is very difficult to get a satisfied effect only by using the conventional recognition method. In this paper, we establish a fuzzy pattern recognition method to vehicle recognition using the two stages. First, we carry out the classification of truck and car by the fuzzy support vector machine theory. Second , we respectively set up recognition model of car and truck by fuzzy neural networks. In order to get a good stability and accuracy model, we used genetic algorithm to optimize membership and reasoning rule at the same time, and then separate the truck and the car.The train data were used to train and modified the model parameters, the test data were used to testify the effect of the recognition model.The result of experimentation testifies the accurateness is 100 percent in the first stage and 93.3 percent in the second stage.